Reviewed by the Help Dementia Editorial Team — our editors review every article for accuracy against guidance from the National Institute on Aging, the Alzheimer’s Association, and peer-reviewed sources.
Yes, new tests can still miss non-Alzheimer’s dementias, even as technology advances. While recent AI-powered blood tests show impressive 92.3% accuracy in distinguishing between dementia types, the headline-grabbing accuracy masks a critical reality: these tests are still developed primarily in research settings, not yet widely available in most clinical practices, and they can still fail to catch overlapping conditions that patients actually live with. More importantly, they remain unavailable to the vast majority of people who need them right now. A 62-year-old man in rural Iowa goes to his doctor with memory problems and tremors.
His physician tests him for Alzheimer’s disease, gets a positive result, and prescribes dementia medications based on that diagnosis alone. Two years later, an autopsy reveals he had both Alzheimer’s disease and Lewy body disease—a combination that would have called for entirely different treatment and symptom management. This scenario isn’t uncommon. Research shows that over 61% of people clinically diagnosed with Alzheimer’s disease actually have autopsy-confirmed Lewy body disease they never knew about. This is the diagnostic gap that new tests promise to close, but whether they can do so in real-world practice remains an open question.
Table of Contents
- What Makes the New AI Blood Test Different—And Where It Still Falls Short
- The Persistent Problem of Misdiagnosis in Dementia Care
- The Particular Challenge of Lewy Body Disease and Other Non-Alzheimer’s Conditions
- The Gap Between Research and Clinical Reality
- Why Overlapping Conditions Make Testing Even More Complex
- The Role of Specialist Access and Clinical Training
- Looking Forward—Progress With Caveats
- Conclusion
What Makes the New AI Blood Test Different—And Where It Still Falls Short
The new AI-powered blood test represents a significant leap forward. Researchers at Washington University School of Medicine trained an artificial intelligence system on blood protein data from over 3,200 individuals with various dementia diagnoses and healthy controls, then verified the results on an independent group of 225 people who underwent autopsy confirmation. The system identifies 15 specific protein biomarkers and can detect when multiple disease processes are present in the same patient—a major advantage over traditional testing that often looks for one disease at a time.
But accuracy in a research study and accuracy in a community clinic are two different things. The 92.3% figure comes from carefully controlled conditions, with blood samples processed in specialized laboratories and analyzed by researchers who know exactly what they’re looking for. In the real world, samples may be handled differently, labs may lack the required equipment, and clinicians may not have the training to order these tests or interpret results correctly. Additionally, the test was developed using autopsy-confirmed diagnoses, which means it’s calibrated against people who died—a population that may not perfectly represent living patients being evaluated for early cognitive symptoms.
Why New Tests Still Leave Misdiagnosis Problems in Dementia Care
The challenge of missing non-Alzheimer’s dementias isn’t new, and it runs deeper than diagnostic technology alone. Dementia remains undetected in healthcare settings for more than 50% of people who have it. Many never receive accurate diagnoses because different dementia diseases are genuinely difficult to distinguish, symptoms overlap significantly, and they frequently occur together in the same person. A patient with early frontotemporal dementia may present with personality changes that a busy primary care doctor might attribute to depression or stress. Someone with Lewy body disease might be misdiagnosed with Parkinson’s disease because tremors and movement problems are so prominent.
By the time a specialist sees the patient, important windows for early intervention have closed. The statistics are sobering. Among participants with autopsy-confirmed coexisting Alzheimer’s disease and Lewy body disease, over 61% were clinically diagnosed as having Alzheimer’s disease only—the Lewy pathology was completely missed during life. Even more striking, among those clinically diagnosed with Alzheimer’s disease, over 32% had concurrent Lewy body neuropathology at autopsy. These aren’t isolated cases. Experts recognize that barriers to timely and accurate dementia diagnosis are multifactorial, involving gaps in clinical training, limited access to specialists, and the sheer complexity of distinguishing between conditions that can look very similar to the untrained eye.
The Particular Challenge of Lewy Body Disease and Other Non-Alzheimer’s Conditions
Dementia with Lewy bodies presents a unique problem for new tests. Unlike Alzheimer’s disease, which has well-established blood biomarkers, direct laboratory-measurable biomarkers for Lewy body disease are not currently known. Diagnosis instead relies on indirect methods: looking for reduced dopamine transporter uptake on PET or SPECT imaging, abnormalities on myocardial scintigraphy, or characteristic patterns on REM sleep polysomnography. These imaging and sleep studies are expensive, time-consuming, and not available in all communities.
The new AI blood test does attempt to distinguish Lewy body disease from other conditions, but here’s where limitations become obvious: blood biomarkers achieved AUC (area under the curve) values of only 74%–76% for differentiating Alzheimer’s or healthy controls from dementia with Lewy bodies. That’s lower than accuracy for frontotemporal dementia, which reached 87%–89%. In practical terms, this means the test is more likely to miss Lewy body disease or misidentify it as something else. A woman in her 70s with hallucinations, parkinsonian features, and cognitive decline gets a blood test that suggests Alzheimer’s disease, but she actually has Lewy body disease—a condition that often responds poorly to some Alzheimer’s medications and sometimes requires different treatment approaches entirely.
The Gap Between Research and Clinical Reality
While biomarkers are progressing rapidly in research settings and specialized dementia clinics, validated markers for routine clinical use remain limited for many dementia subtypes. This gap matters because new tests often take years to move from publication to widespread availability. A blood test published in a major journal today might not be available in most hospitals or primary care practices for five years or more. Insurance may not cover it. Laboratories may lack the specialized equipment needed to run the test. Clinicians in rural areas or smaller cities may not even know the test exists.
Consider the practical reality of a patient journey: A 58-year-old woman notices increasing difficulty with language and personality changes. Her family doctor runs standard cognitive tests and refers her to a neurologist. If she lives near a major medical center, she might get seen in a few months. If she lives in a less urban area, it could be a year. By the time she might access the new AI blood test, her cognitive decline has progressed. Meanwhile, the months or years of diagnostic delay have been months or years without access to treatments, support services, and lifestyle modifications that might have helped. The test doesn’t exist in a vacuum—it exists in a healthcare system where timing and access shape outcomes just as much as accuracy does.
Why Overlapping Conditions Make Testing Even More Complex
The human brain often shows pathological evidence of multiple dementia diseases simultaneously. This comorbidity is the rule, not the exception. Someone might have Alzheimer’s disease pathology, Lewy body pathology, and vascular dementia all at the same time. The new AI test is designed to identify when multiple disease processes are present, which is valuable—but it still faces the challenge of accurately characterizing each component when they coexist.
Imagine a 75-year-old man with mixed dementia: Alzheimer’s disease and vascular changes from a history of high blood pressure. The blood test might identify both conditions, which is excellent. But which one is causing which symptom? Which one should treatment prioritize? Is cognitive training going to help if vascular problems are the main driver? These are questions that require clinical judgment beyond what even an accurate test can provide. The test gives you information; doctors still have to interpret it and apply it to a specific person’s life. That interpretation process is where some of the highest risks of missed diagnoses still occur.
The Role of Specialist Access and Clinical Training
Getting an accurate diagnosis increasingly depends on seeing the right specialist—but specialists in dementia are concentrated in urban centers and academic medical centers. A primary care physician in a small town may never see a dementia specialist in their career. They rely on standard diagnostic protocols and their own training, which may be years out of date. Even when new tests become available, clinicians need education about what the test does, what its limitations are, and how to discuss results with patients and families.
A patient’s experience might go like this: They finally get access to the new AI blood test through their neurologist, and it shows they have dementia with Lewy bodies. But the neurologist is trained primarily in Alzheimer’s disease and may not be familiar with the best current evidence-based approaches to managing Lewy body disease, including which medications are most appropriate and which should be avoided. The patient then needs to travel to a major medical center for a second opinion, or their physician has to engage in continuing education to properly manage the condition. Access to the test is only the beginning; access to the expertise to use it well is another matter entirely.
Looking Forward—Progress With Caveats
The development of more accurate diagnostic tests is undoubtedly progress. The Washington University research and similar efforts are moving the field toward better identification of dementia subtypes. But progress in diagnostics doesn’t automatically translate to progress in patient outcomes unless it’s coupled with broader changes: wider availability of testing, training programs for clinicians, development of better treatments for non-Alzheimer’s dementias, and insurance policies that cover these tests.
The next decade will likely bring even more sophisticated blood tests and perhaps new biomarkers that better distinguish dementia types. But the question “Could new tests miss non-Alzheimer’s dementias?” will still have a nuanced answer: The tests themselves may become more accurate, but missed diagnoses will continue as long as people lack access to the tests, as long as clinicians aren’t trained to order them or interpret them, and as long as dementia remains a low priority in primary care. The future of dementia diagnosis depends on more than better science—it depends on better systems.
Conclusion
New diagnostic tests represent genuine progress in dementia care, offering greater accuracy in distinguishing between Alzheimer’s disease, Lewy body disease, frontotemporal dementia, and other conditions. A 92.3% accuracy rate in identifying multiple disease processes is genuinely meaningful for patients and families seeking answers. But accuracy in research settings is only part of the equation. The larger challenge is ensuring these tests reach patients who need them, that clinicians are trained to use them effectively, and that diagnoses lead to appropriate treatment and support.
If you or a loved one is experiencing cognitive changes, don’t wait for the perfect test to become available in your community. Work with your doctor to get evaluated now using the tools and expertise available to you. Ask your physician whether you might be referred to a neurologist or dementia specialist. Even without the newest blood tests, careful clinical evaluation, imaging when appropriate, and thoughtful assessment can lead to accurate diagnoses that make a real difference in care planning and quality of life.
You Might Also Like
- Why Doctors Need Clear Guidelines for New Dementia Tests
- Can Early Parkinson’s Detection Research Help Dementia Testing?
- Could Precision Medicine Change Dementia Care?
Related reading
- can Press Releases Make Alzheimer’s Discoveries Sound Too Certain
- how to Tell If a Dementia Breakthrough Is Real
- why Families Should Watch for Words Like Experimental and Early
- what Alzheimer’s Headlines Often Leave Out
- can News About Dementia Research Give False Hope
For more on this topic, see CDC — Alzheimer’s and Dementia.
